diff --git a/db/message_storage.py b/db/message_storage.py
index 56d20ab..326b130 100644
--- a/db/message_storage.py
+++ b/db/message_storage.py
@@ -43,6 +43,24 @@ class MessageStorageDB(BaseDBOperator):
         params = (hours_ago, group_id, min_content_length)
         return self.execute_query(sql, params) or []
 
+    def get_latest_image_message(self, group_id: str, before_timestamp: str = "", hours_ago: int = 8) -> Optional[Dict]:
+        """获取指定群最近一条已落盘图片消息"""
+        sql = """
+        SELECT timestamp, sender, content, message_type, image_path
+        FROM messages
+        WHERE timestamp >= DATE_SUB(NOW(), INTERVAL %s HOUR)
+          AND group_id = %s
+          AND message_type = 3
+          AND image_path IS NOT NULL
+          AND image_path <> ''
+        """
+        params: List = [hours_ago, group_id]
+        if before_timestamp:
+            sql += " AND timestamp <= %s"
+            params.append(before_timestamp)
+        sql += " ORDER BY timestamp DESC LIMIT 1"
+        return self.execute_query(sql, tuple(params), fetch_one=True)
+
     def get_member_recent_messages(self, group_id: str, wxid: str, days: int = 30,
                                    limit: int = 200, include_today: bool = True) -> List[Dict]:
         """获取指定群成员近期消息"""
diff --git a/plugins/ai_auto_response/context_builder.py b/plugins/ai_auto_response/context_builder.py
index 5ad747d..d1f5093 100644
--- a/plugins/ai_auto_response/context_builder.py
+++ b/plugins/ai_auto_response/context_builder.py
@@ -22,6 +22,7 @@ class ContextBuilder:
         reply_mode: str,
         vector_memories: List[Dict],
         quote_context: Dict | None = None,
+        image_context: Dict | None = None,
     ) -> Dict:
         recent_lines = []
         for item in recent_messages[-self.recent_context_size:]:
@@ -45,6 +46,7 @@ class ContextBuilder:
             "vector_memory_prompt": self._build_vector_memory_prompt(vector_memories),
             "group_profile_prompt": self._build_group_profile_prompt(group_profile or {}),
             "quote_prompt": self._build_quote_prompt(quote_context or {}),
+            "image_prompt": self._build_image_prompt(image_context or {}),
             "current_message": f"{sender_name}: {content}",
         }
 
@@ -136,3 +138,14 @@ class ContextBuilder:
             f"被引用内容：{quote_body}" if quote_body else "",
         ]
         return "\n".join([line for line in lines if line])
+
+    @staticmethod
+    def _build_image_prompt(image_context: Dict) -> str:
+        if not image_context:
+            return ""
+        lines = [
+            "已附带最近一张群图片作为上下文。",
+            f"图片发送者：{image_context.get('sender_name', '未知成员')}",
+            f"图片说明：{image_context.get('hint', '')}" if image_context.get("hint") else "",
+        ]
+        return "\n".join([line for line in lines if line])
diff --git a/plugins/ai_auto_response/main.py b/plugins/ai_auto_response/main.py
index ba4c5e6..0e095d4 100644
--- a/plugins/ai_auto_response/main.py
+++ b/plugins/ai_auto_response/main.py
@@ -241,7 +241,12 @@ class AIAutoResponsePlugin(MessagePluginInterface):
         vector_memories = []
         if self.vector_memory.should_search(reply_mode, trigger.trigger_type, memory_hints.get("returning_member_state", "")):
             vector_memories = self.vector_memory.search(content, room_id, sender)
+        image_context = self._build_recent_image_context(message, room_id, content, quote_context)
         image_urls = await self._prepare_quote_image_inputs(bot, quote_context)
+        if not image_urls and image_context:
+            recent_image_url = self._build_local_image_data_url(str(image_context.get("image_path", "") or ""))
+            if recent_image_url:
+                image_urls = [recent_image_url]
         self._log_event(
             "context",
             room_id=room_id,
@@ -268,6 +273,7 @@ class AIAutoResponsePlugin(MessagePluginInterface):
             reply_mode=reply_mode,
             vector_memories=vector_memories,
             quote_context=quote_context | {"has_image_attachment": bool(image_urls)},
+            image_context=image_context,
         )
 
         system_prompt = self.persona_engine.build_system_prompt(group_profile)
@@ -377,6 +383,7 @@ class AIAutoResponsePlugin(MessagePluginInterface):
             f"当前群聊消息：\n{recent_text}\n\n"
             f"当前发言：{context.get('current_message', '')}\n"
             f"引用补充：\n{context.get('quote_prompt', '') or '无'}\n"
+            f"图片补充：\n{context.get('image_prompt', '') or '无'}\n"
             f"触发类型：{context.get('trigger_type', 'none')}\n"
             f"回复模式：{context.get('reply_mode', 'social_short')}\n"
             f"当前心流状态：{context.get('flow_state', 'idle')}\n"
@@ -725,6 +732,40 @@ class AIAutoResponsePlugin(MessagePluginInterface):
             return title[:220].strip()
         return ref_content[:220].strip()
 
+    def _build_recent_image_context(
+        self,
+        message: Dict[str, Any],
+        room_id: str,
+        content: str,
+        quote_context: Dict[str, str],
+    ) -> Dict[str, str]:
+        if quote_context:
+            return {}
+        if not self._is_recent_image_followup(content):
+            return {}
+        latest_image = self.memory_store.get_latest_image_message(
+            room_id,
+            before_timestamp=str(message.get("timestamp") or ""),
+        )
+        if not latest_image:
+            return {}
+        sender = str(latest_image.get("sender", "") or "")
+        sender_name = self._get_sender_name(room_id, sender) if sender else "未知成员"
+        return {
+            "sender_name": sender_name,
+            "image_path": str(latest_image.get("image_path", "") or ""),
+            "hint": "用户当前这句大概率是在追问这张最近图片",
+        }
+
+    @staticmethod
+    def _is_recent_image_followup(content: str) -> bool:
+        text = str(content or "").strip().lower()
+        if not text:
+            return False
+        image_words = ["图", "图片", "照片", "截图"]
+        ask_words = ["看看", "看下", "帮我看", "帮看看", "这个", "咋样", "什么", "识别", "分析"]
+        return any(word in text for word in image_words) and any(word in text for word in ask_words)
+
     async def _prepare_quote_image_inputs(self, bot: WechatAPIClient, quote_context: Dict[str, str]) -> List[str]:
         if not quote_context or quote_context.get("quote_type_label") != "引用图片":
             return []
@@ -746,6 +787,21 @@ class AIAutoResponsePlugin(MessagePluginInterface):
             return []
         return [data_url]
 
+    def _build_local_image_data_url(self, image_path: str) -> str:
+        if not image_path:
+            return ""
+        relative_path = image_path.lstrip("/\\").replace("/", "\\")
+        full_path = self.get_main_path() / relative_path
+        if not full_path.exists():
+            return ""
+        try:
+            image_bytes = full_path.read_bytes()
+        except Exception:
+            return ""
+        image_type = imghdr.what(None, h=image_bytes) or "jpeg"
+        raw_base64 = base64.b64encode(image_bytes).decode("utf-8")
+        return f"data:image/{image_type};base64,{raw_base64}"
+
     @staticmethod
     def _extract_quote_image_info(ref_content: str) -> Dict[str, str]:
         if not ref_content:
diff --git a/plugins/ai_auto_response/memory_store.py b/plugins/ai_auto_response/memory_store.py
index 4c0b248..16fa94b 100644
--- a/plugins/ai_auto_response/memory_store.py
+++ b/plugins/ai_auto_response/memory_store.py
@@ -20,6 +20,10 @@ class MemoryStore:
         size = int(self.config.get("recent_context_size", 30))
         return recent[-size:]
 
+    def get_latest_image_message(self, room_id: str, before_timestamp: str = "") -> Optional[Dict]:
+        hours = int(self.config.get("active_context_hours", 8))
+        return self.message_db.get_latest_image_message(room_id, before_timestamp=before_timestamp, hours_ago=hours)
+
     def get_member_context(self, room_id: str, wxid: str) -> Optional[Dict]:
         if not self.config.get("enable_member_context", True):
             return None